This technology, BeatProfiler, is a desktop graphical-user-interface (GUI) that leverages deep learning for high-throughput, streamlined analysis of in vitro human cardiac cells and engineered tissues, primarily aimed at accelerating drug development and disease phenotyping in cardiovascular research.

Unmet Need: Stand-alone and user-friendly multimodal cardiac model analysis tool

The current gold standard for analyzing cardiac models involves labor-intensive image analysis techniques that often require additional coding and the use of external packages like MATLAB or ImageJ. These methods can be cumbersome and time-consuming, posing significant challenges for researchers studying cardiovascular disease. Furthermore, the complexity of these techniques may limit their accessibility and reproducibility to researchers without extensive programming knowledge. There is a pressing need for a more intuitive tool that can automate quantitative analysis of cardiac models, thereby accelerating the process of drug development and disease classification.

The Technology: Multimodal analysis tool for cardiac tissue images

This technology is a desktop graphical-user-interface (GUI) designed for both Windows and macOS that uses deep learning to automate and streamline the analysis of in vitro cardiac models called BeatProfiler. It supports over five video formats and can read videos of single cells, 2D monolayers, 3D spheroids, 3D tissues, and other models. The software extracts multimodal features such as contractility, calcium dynamics, and tissue force output, providing a comprehensive analysis of cardiac tissue. Unlike current options, BeatProfiler is intuitive and standalone, eliminating the need for additional coding or external packages.

This technology has been validated using videos of in vitro cardiac tissues.

Applications:

High-throughput drug screening
Disease modeling for various cardiovascular conditions
Precision medicine diagnostics for personalized treatment strategies
Research tool for studying cardiac tissue behavior and response
Educational tool for teaching and understanding cardiac tissue function

Advantages:

Streamlined, objective and automated analysis of cardiac models
User-friendly interface that does not require additional coding or external packages
Supports multiple video formats and cardiac modalities
Utilizes deep learning for precise and reproducible extraction of multimodal features including contractility, calcium dynamics, and force generation in engineered tissues
Standalone platform, eliminating the need for other software or tools

Lead Inventors:

Gordana Vunjak-Novakovic, Ph.D.

Youngbin Kim

Related Publications:

In preparation

Tech Ventures Reference:

Automated software tool for analysis of cardiac tissue function